Granite Vision uses a small model to read complex documents
IBM emphasizes tables, charts, and linked fields.
Command-line tools now have two kinds of users
Mistral’s Spaces engineering story considers interfaces for people and agents.
Cohere Transcribe turns speech into a separate building block
The company expands into downloadable speech recognition.
EVA evaluates the conversation, not just the transcript
ServiceNow researchers look at both task completion and the human experience of voice AI.
Voxtral adds a voice of its own
Mistral introduces text-to-speech for conversational and other audio workflows.
Measuring intelligence takes more than one score
Google DeepMind proposes a cognitive framework for evaluating AI progress.
GPT-5.4 mini and nano find roles inside larger workflows
OpenAI expands its smaller-model options for fast and repeated tasks.
Mistral Small 4 combines reasoning, vision, and code
The release brings several strands of Mistral’s work into one model.
Yandex brings a stronger Alice model to smart devices
The update broadens access beyond the previously required add-on.
Implementing AI becomes a skill in its own right
Anthropic launches a partner network for adoption and technical support.
A Rails testing agent needs to run what it writes
Mistral engineers describe a specialized loop for generating and checking tests.
LeRobot 0.5 connects more of the robotics workflow
Hugging Face expands hardware, control policies, and data tooling.
GPT-5.4 combines code, documents, and computer use
OpenAI pushes its general model toward completed professional tasks.
Modular Diffusers makes generation a set of building blocks
Hugging Face offers a more composable way to build visual pipelines.
Gemini 3.1 Flash-Lite targets high-volume work
Google introduces a smaller option for frequently repeated tasks.